Data Extractivism as the Foundation of the Advertising Business
Free services were never free. Data is the raw material of the ad economy – and we are the ones supplying it.
Data Extractivism as the Foundation of the Advertising Business
When we conquered the internet in the early 2000s, everything felt like freedom. Information suddenly became accessible to everyone. Knowledge was only a few clicks away. Search engines made libraries seem obsolete, social networks brought people together across continents, and every new digital service seemed to turn another piece of the future into reality.
Looking back, though, we were remarkably naive.
We believed free services were actually free. We enjoyed the new possibilities without questioning the real price. None of us read the terms of service. Nobody seriously asked why billion-dollar companies were willing to give us search engines, email, social networks, cloud storage and smartphone operating systems at no charge. We know the answer today: We were never the customers. We were always the suppliers of raw material.
Human experience as a resource
This is exactly where Ingo Dachwitz and Sven Hilbig start in their book Digitaler Kolonialismus (Digital Colonialism). The authors describe data as the central raw material of a new global economy. Unlike classical colonialism, today's extraction does not target gold mines, plantations or oil fields. The resource is human experience. Every search query, every location, every click, every photo, every movement with a smartphone and every interaction on social networks becomes a dataset that can be collected, analysed and turned into profit.
That is why the term data extractivism hits the core of the problem. Just as earlier colonial powers extracted resources from their colonies and let the profits flow into the centres of power, data is collected worldwide today while most of the economic value ends up with a handful of technology corporations. The global imbalance is obvious. The data comes from billions of people around the planet. The profits land mostly in the headquarters of American tech companies.
Who owns the infrastructure
Google, Apple, Meta, Amazon and Microsoft, often grouped as GAFAM, now control central infrastructures of our digital lives. They run search engines, social networks, cloud platforms, app stores, ad networks and, increasingly, the infrastructure for artificial intelligence. That gives them a position of power that is historically unprecedented in this form.
Social networks such as Instagram and TikTok have pushed this further. Their business is not producing content. Their business is maximising attention. The platforms compete for their users' limited time. Every extra minute means new data points. Every new data point makes the models more precise. And every more precise model raises the value of the ad inventory.
The real raw material of the digital economy is therefore not technology. It is human behaviour.
What will you do next?
The longer users stay on a platform, the more accurately their behaviour can be predicted. The advertising industry is no longer interested only in who we are today. It wants to know who we will be tomorrow. Are we about to buy a car? Expecting a child? About to move house? Interested in financial products? Likely to switch health insurers? The commercially valuable question is no longer: "Who are you?" The decisive question is: "What will you do next?"
Seen from that angle, online advertising suddenly looks different. The most profitable business model of the internet economy is not advertising itself. It is an information advantage. Targeted advertising only works because companies know more about us than we may know about our own future behaviour. Data is collected, aggregated, enriched, categorised and finally sold through complex ad platforms.
The invisible data industry
Companies such as Google, Meta or Amazon often sit at the centre of public attention, but behind the scenes there is an entire industry of data brokers and profile vendors. Firms such as Acxiom, Experian, Nielsen, Criteo or, formerly, Oracle Data Cloud have spent years building gigantic data holdings. Acxiom talks about billions of records and identities. Experian holds information on around 1.4 billion people worldwide. Nielsen advertises a reach that covers a large part of the world's population. The sheer size of these collections makes clear that this is no longer classic market research. It is industrialised behavioural analysis at global scale.
Especially striking was the investigation by netzpolitik.org and The Markup, who gained access to datasets from the advertising platform Xandr, which now belongs to Microsoft. The journalists found around 650,000 audience categories. Even that number is hard to grasp. It shows how finely people are categorised today. The digital advertising industry is no longer trying to describe target groups. It is trying to break individuals into as many commercially usable attributes as possible.
Invisible costs
At the same time, a striking double standard has taken hold. We debate the ecological footprint of planes, cars and industrial plants at length. That matters. But we talk astonishingly little about the ecological costs of our digital infrastructure. It would of course be absurd to claim that an AI-generated video "used 20 trees and 1,000 litres of water". That wording is deliberately exaggerated and sarcastic. Still, it points to a real fact: data centres need enormous amounts of energy, cooling and resources. We can see the exhaust of a combustion engine immediately. The environmental impact of digital services often stays invisible.
Perhaps that is one of the greatest achievements of the platform economy: it made surveillance invisible. People in the 1980s and 1990s would probably have resisted many forms of today's data collection. Visible walls, border fences or checkpoints produce resistance. Invisible data collection produces convenience. Whoever wraps surveillance in a free app and pairs it with entertainment no longer has to force anyone.
The bill arrived later
The most uncomfortable insight, though, is that we ourselves were part of this development. Generation X and the millennials helped build and shape the open internet. We used the platforms with enthusiasm. We willingly traded our data for convenience. Nobody forced us. The bill was simply presented years later.
That is why data extractivism is not a side issue of the digital economy. It is its foundation. The data economy does not work despite the collection of personal information. It works because of it. As long as attention remains the internet's most important currency, data will remain its most important resource. The central power question of the digital age is therefore not who builds the best artificial intelligence or the fastest processor. The decisive question is who controls the data.
Because in the end, the same rule applies as in every resource boom in history:
Whoever controls the resource controls the market.
And in the digital age, the most important resource is not oil, gold or lithium.
It is you.
Sources
[1] Ingo Dachwitz / Sven Hilbig: Digitaler Kolonialismus. Wie Tech-Konzerne und Großmächte die Welt unter sich aufteilen. C.H. Beck, Munich 2025. ISBN 978-3-406-82302-2.
https://www.beck-shop.de/dachwitz-hilbig-digitaler-kolonialismus/product/37000393
[2] Acxiom Real ID: claims of around 2.6 billion identities and 260 million U.S. consumers.
https://www.acxiom.com/products/real-id/
[3] Experian Investor Roadshow (2023): "data on 1.4 billion people and 191 million businesses".
https://www.experian.com/content/dam/marketing/global/plc/en/assets/documents/results-and-presentations/2023/experian-ir-roadshow-deck-jan-to-mar-2023.pdf
[4] Nielsen: company description stating services covering more than 90 percent of the world's population and GDP.
https://www.adobe-audience-finder.com/data_partner/nielsen/
[5] Criteo: registered data broker (including California and Texas), ad profiles based on user behaviour.
https://www.criteo.com/privacy/
[6] Digiday: Oracle shut down its advertising business, including Oracle Data Cloud, in 2024.
https://digiday.com/marketing/the-rise-stall-and-fall-of-oracles-advertising-business/
[7] AT&T / Microsoft: sale of Xandr to Microsoft, completed in 2022.
https://www.prnewswire.com/news-releases/att-agrees-to-microsoft-acquisition-of-xandr-301448996.html
[8] netzpolitik.org: analysis of the Xandr marketplace list with more than 650,000 audience categories.
https://netzpolitik.org/2023/microsofts-datenmarktplatz-xandr-das-sind-650-000-kategorien-in-die-uns-die-online-werbeindustrie-einsortiert/
[9] The Markup: joint investigation into 650,000 audience segments on the Xandr platform.
https://themarkup.org/privacy/2023/06/08/from-heavy-purchasers-of-pregnancy-tests-to-the-depression-prone-we-found-650000-ways-advertisers-label-you
[10] IEA: data centres used around 415 TWh of electricity in 2024, about 1.5 percent of global electricity consumption.
https://www.iea.org/reports/energy-and-ai/executive-summary